AI & ChatGPT Mini Crash Course - Eliminate workplace burnout & save 16+ hours every week. Learn 20+ AI tools, prompting techniques & hacks for free.

Save your seat here (first 100 people only) ⏰

Today’s top AI Highlights:

1. Move over Upwork, AI Agents are hiring humans (and paying them)

2. Opensource model distillation kit for creating small AI models

3. Fine-tune Mistral AI models on your data without any coding

4. OpenAI invests $60M in Opal; plans to make AI-powered webcams

5. AI that knows exactly when you need to work or take a break

& so much more!

Read time: 3 mins

Latest Developments

AI That Pays Humans 💰

Imagine an AI agent acting as your Chief Marketing Officer, devising marketing strategies and tasks, and delegating these tasks to humans. This AI agent has the added ability to pay humans to execute those tasks. That's exactly what Payman, a new AI platform, makes possible. They have secured $3 million in pre-seed funding from investors like Visa and Coinbase Ventures, to build AI-human payments infrastructure and develop robust collaboration between humans and AI.

Key Highlights:

1. AI Takes Charge - Developers can create AI agents to act like mini-project managers. For instance, you create an AI agent to promote your company on Twitter, set a budget of $200, and provide some background information on the company. The AI agent, leveraging Payman's platform, then creates a detailed task outlining the requirements for crafting engaging tweets about the company.

2. Humans Get Paid - This task is published on Payman's marketplace, visible to users looking to earn money by completing tasks. A user sees the task, accepts it, creates the tweets, submits their work, and receives the $200 payment upon approval – all facilitated through Payman.

3. Seamless and Secure - Payman handles the entire payment process, ensuring smooth transactions between AI agents and humans.

4. Transparency - Payman offers transparency for both developers and humans. Developers can track their agent's spending and task completion, while humans can browse available tasks, clearly see payment details, and manage their earnings directly through the platform.

Opensource Toolkit for Compact and Powerful AI 🌟

Arcee AI has released DistillKit, a new open-source tool for creating smaller, more efficient AI models. The aim is to create small AI models that are competitive with their larger counterparts but efficient enough to run on devices like laptops and smartphones. This new toolkit uses "model distillation" to teach smaller models (students) the knowledge of larger, more resource-intensive models (teachers). The two techniques on which this toolkit is built help these smaller models learn not just answers, but the reasoning behind them.

Key Highlights:

1. Logit-based Distillation - This method is like having the big model show its work to the smaller model. The smaller AI doesn't just learn the right answers – it also learns how confident the big model is about different possible answers. This helps the smaller model think more like the bigger one.

2. Hidden States-based Distillation - This approach teaches the smaller model to understand information in a similar way to the big model. It helps the smaller model learn the thinking process, not just the final answers.

3. Open Source - DistillKit includes features like supervised fine-tuning with Distillation, along with experimental results and documentation. You can readily use and adapt the provided code and methods for your own model distillation projects.

Develop and Share AI Applications Without Coding 👨‍💻

Mistral AI is making it simpler for developers to customize and deploy LLMs with new fine-tuning capabilities on its La Plateforme and the alpha release of Agents. You can now fine-tune Mistral AI’s flagship models like Mistral Large 2 and Codestral directly on their platform without any coding. Agents offer a powerful framework for building and sharing AI agents by giving simple instructions in natural language. You can leverage the power of Mistral's models with a lower barrier to entry, creating highly tailored AI solutions.

Key Highlights:

Fine-tuning

1. No-Code Customization - Mistral's La Plateforme offers a no-code interface for fine-tuning, allowing developers to customize models using base prompts, few-shot learning, or full fine-tuning with custom datasets.

2. Memory-Efficient Training - You can use the mistral-finetune opensource library that fine-tunes by training only a small fraction of the model's weights, making the process faster and less resource-intensive.

Agents

3. Streamlined Workflow - Agents framework wraps AI models with additional context and instruction to create powerful AI agents. Just select the AI model, and give it specific instructions and examples in simple language to create custom behaviour and workflows.

4. Sharing - These custom AI agents can be shared easily within organizations, promoting collaboration and code reuse. They can be exposed via Le Chat or API.

Quick Bites

1. OpenAI is investing $60 million in Opal Camera Inc., a startup that makes high-end webcams like the Tadpole. Opal will continue making webcams after closing the round but will shift some resources to developing AI-powered devices that can be used as “creative tools” by consumers.
   

2. Reddit is testing AI-generated summaries at the top of search results to help users find and explore content more easily. Reddit had entered into a partnership with OpenAI in May to use its technology and give access to Reddit data. It has a similar agreement with Google as well.
   

3. Amazon-owned audiobook company Audible is testing an AI-powered search feature called Maven to help users find audiobooks using natural language queries. It is now available to select U.S. customers on iOS and Android.
   

4. Apple had announced that it is not releasing any Apple Intelligence feature in the EU this year due to regulatory constraints. However, the recent beta release notes indicate Apple Intelligence may become available for EU users on the Mac, but not on iOS and iPadOS 18.

Tools of the Trade

1. Stormy: Uses AI to help you stay focused by offering personalized interventions whenever you get distracted. It understands your work patterns and tailors its suggestions to keep you on track and productive.

2. GPT Researcher: Your AI mate for thorough and speedy online research, handling everything from gathering sources to organizing results. It leverages multi-agent frameworks for parallel processing to produce detailed, unbiased research reports on any topic.

3. RAG Me Up: Use RAG on your own data with an easy-to-use server and customizable UIs. It works best on a GPU and offers flexible, code-free configuration for indexing and querying your documents.

4. Awesome LLM Apps: Build awesome LLM apps using RAG for interacting with data sources like GitHub, Gmail, PDFs, and YouTube videos through simple texts. These apps will let you retrieve information, engage in chat, and extract insights directly from content on these platforms.

Hot Takes

1. I think it's really underrated the battlefield shaping Zuck did by open sourcing Meta's AI work
   It not only changed OpenAI's entire flow, but pushed Anthropic into a more product centric strategy (as opposed to research centric) - Mike Krieger hire exemplifies this ~
   Reggie James
   

2. Sama, if you want to stop bleeding talent to Anthropic, you need to release GPT-5 now and prove OpenAI is the future. ~
   Flowers

Meme of the Day

That’s all for today! See you tomorrow with more such AI-filled content.

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!

PS: We curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!